Internet for Pharmaceutical and Biotech Communities
| Newsletter | Advertising |



Outsourcing Guide

Cont. Education


Training Courses

Web Seminars


Buyer's Guide

Home Page

Pharm Patents /

Pharm News

Federal Register

Pharm Stocks

FDA Links

FDA Warning Letters


Pharm/Biotech Events


Advertiser Info

Newsletter Subscription

Web Links


Site Map



Title:  Designing compounds specifically inhibiting ribonucleic acid function
United States Patent: 
October 2, 2007

Schimmel; Paul R. (Cambridge, MA)
Massachusetts Institute of Technology (Cambridge, MA)
Appl. No.: 
May 26, 1994


Web Seminars -- Pharm/Biotech/etc.


A method for designing compounds specifically targeting RNA sequences, based on the discovery of short, specific sequences within RNA that are critical to function, using modeling of the compound to effect binding to the nucleotide sequences in the RNA in combination with secondary and/or tertiary structure associated with the minor groove of the RNA in the region of the critical sequences. In the preferred method, computer modeling is used along with analysis of the targeted RNA sequence to design molecules binding to the targeted RNA by covalent or hydrogen binding. Appropriate molecules are synthesized using known methodology that have the required structure and chemical characteristics to specifically bind the critical region of the RNA and thereby inhibit the function of the RNA. Molecules known to bind to RNA can also be modified using this method to increase specificity, and thereby decrease toxicity.


A method for designing compounds specifically targeting RNA sequences by modeling the compound to bind to crucial short nucleotide sequences in the RNA in combination with secondary and/or tertiary structure associated with the minor groove of the RNA is provided. In the preferred method, a critical sequence of the RNA is identified, then computer modeling is used in combination with analysis of the targeted RNA sequence to design molecules binding to the targeted RNA by covalent or hydrogen binding. Appropriate molecules specifically inhibiting the function of the targeted RNA are synthesized using known methodology that have the required secondary structure and chemical characteristics. Molecules known to bind to RNA can also be modified using this method to increase specificity, and thereby decrease toxicity.

Much of the design of these compounds, as well as the inhibitory effect of these compounds, is based on studies on the recognition of RNAs by proteins in combination with in vitro RNA synthesis. For example, studies have demonstrated that the G3:U70 base pair of tRNA.sup.Ala is critical for its function. By taking advantage of sequence differences around G3:U70 between the human tRNA.sup.Ala and that of a pathogenic organism, selective drug binding can be achieved and protein synthesis by the pathogenic organism inhibited. Another example involves the interaction between the RNA-dependent reverse transcriptase of retroviruses and the specific tRNA that acts as a primer for reverse transcriptase. The annealing of the primer tRNA to the primer binding site is the first step in initiation of cDNA synthesis by reverse transcriptase, and thus represents a potential target for the arrest of viral multiplication. This can also be used as an assay for testing inhibitors of the binding reaction, for example, using glycerol gradient centrifugation to detect the presence of a complex between HIV reverse transcriptase and primer lysine tRNA.

Examples demonstrate the targeting of compounds to viral RNA which inhibit viral infection and/or replication, synthesis of compounds inhibiting viral reverse transcriptase, targeting of compounds to bacterial but not eukaryotic tRNA molecules to inhibit bacterial replication, and modification of compounds inhibiting rRNA to impart greater specificity and thereby decrease toxicity to normal cells.


The method of the present invention depends on an understanding of the primary, secondary and tertiary structure of RNA and the determination of short nucleotide sequences essential for functioning of the RNA, particularly specialized RNA such as tRNA and rRNA. Compounds must bind to the targeted RNA with specificity, as determined by the secondary and/or tertiary structure and bonding with associated nucleotides, and effectively, by blocking access to the critical nucleotides required for function of the RNA.

Targeted RNA Molecules.

RNA molecules that can be inhibited include mRNA, tRNA, rRNA, and viral RNA. Both single strand and double strand RNA can be bound and therefore inhibited. Inhibition, as used herein, refers to a decrease in the RNA's function, where function may be transcription, translation, attachment of amino acids, activation of subsequent amino acids as required to form peptides, binding of initiation, elongation and termination factors, peptide bond formation, and translocation.

Proteins that participate in gene regulation, DNA synthesis, and other processes make the majority of their sequence specific contacts with B-form DNA through major groove interactions. Because the deep groove is too narrow for a protein in the form of an alpha-helix to make direct sequence-specific contact, the primary basis for sequence discrimination in RNA is usually the minor groove. Based on the three-dimensional structure of yeast tRNA.sup.Phe, it has been proposed by Rich and Schimmel, Nucl. Acids Res. 4:1649 (1977, that sequence-specific recognition of the tRNA by aminoacyl tRNA synthetases occurs mainly through contacts along the inside surface of the tRNA "L" shape, where the minor groove is available and where the anticodon is located. Recognition and interaction with determinants may occur through hydrogen bonds to either the minor groove exocyclic amino or keto groups, or to the unpaired bases themselves.

Characterization of the Primary, Secondary and Tertiary Structure of the Targeted RNA.

Each RNA is characterized by its primary, secondary, and tertiary structure. Until recently, little has been known about specific RNA sequence and structure. Moreover, there has been a methodological problem in obtaining sufficient quantities of nucleic acid for testing model RNAs for recognition by new compounds. Large scale in vitro and chemical syntheses of RNA is now possible, allowing analysis by x-ray crystallography and other analytical methods. A number of interactive computer graphics programs are also available, which can be used for analysis of secondary and tertiary structure of the RNA. Both of these techniques can be used to predict new and improved molecular compounds that will bind specifically to selected sites on the RNA molecules.

X-ray diffraction analyses have established that virtually all tRNA molecules exist as hydrogen-bonded cloverleaf secondary structures, with tertiary structure formed by additional folding, as depicted schematically in FIG. 2A and by computer modeling in FIG. 2B. High resolution, three-dimensional X-ray structures are available for four tRNAs, showing precise geometries of helical domains and confirming that the stem-loop is precisely folded into an L-shaped three-dimensional conformation, with two helices and major and minor grooves. Pleij, et al., Nucleic Acids Research 13(5), 1717-1731 (1985), reviews the tertiary interaction involving hairpin or interior loops of RNA, and other tertiary structures, and their impact on ribosome function, RNA splicing and recognition of tRNA-like structures.

The amino acid acceptor regions of many viral RNAs have also been sequenced. The secondary structures of viral RNAs probably are not in the form of the typical cloverleaf tRNA. However, the different primary and secondary structures of tRNAs and certain viral RNAs can be recognized efficiently by the same tRNA-specific enzymes, as reviewed by Haenni, et al., Progress in Nucleic Acid Research & Molecular Biology 27:85 (1982). For example, evidence indicates that viral RNAs of high molecular weight and eukaryotic tRNAs similarly recognize aminoacyl-tRNA synthetases. These data support the conclusion that the tertiary, rather than the primary and cloverleaf folding, determines recognition. For example, the brome mosaic plant virus has an RNA that can be specifically tyrosylated by tyrosyl-tRNA synthetases; the core region necessary for aminoacylation has been identified by Dreher and Hall, J. Molec. Biol. 210:41-55 (1988).

Characterization of Critical Sequences within the Targeted RNA Molecule.

Critical sequences of the targeted RNA molecule are determined using a method such as substitutional mutation and comparison of function of the mutated RNA with the original RNA; or base substitution in tRNA and determination of which amino acid is now recognized by the tRNA; the critical sequences may also affect, or be determined by, secondary and tertiary structure of the RNA molecule.

In the first method, substitution mutations are made in the RNA and the function of the mutated RNA compared with that of the original molecule. For example, nucleotide bases in the aminoacyl acceptor region of tRNA can be substituted and the resulting RNA tested to see if (1) an amino acid is attached and (2) if so, which one. The minimum number of nucleotide substitutions (sequence changes) that are required to convert a tRNA from one amino acid identity to another can be determined in this manner.

To practice this method, the RNA molecule is obtained either by in vitro transcription using bacteriophages that encode and synthesize polymerases, such as SP6 and T7, both of which are commercially available, by chemical synthesis, or by isolation from cells that produce the RNA.

A useful approach for elucidating and testing models for recognition is by investigation of substitution mutations of both the protein and of the nucleic acid. For RNA, one obstacle to this approach has been the difficulty in freely generating and isolating mutant and wild-type RNA species from whole cells in quantities that are sufficient for quantitative studies.

In Vitro Synthesis of RNA.

A number of advances in the field of in vitro RNA synthesis have greatly facilitated the generation of sequence and length variants of different RNAs. The in vitro enzymatic synthesis of RNA was originally beset by numerous technical impediments. Initially, transcripts were obtained by use of purified RNA polymerase from E. coli, which has three different subunits in the core enzyme (.alpha., .beta., .beta.') and a separate one for specific initiation (.delta.). Each of these subunits must be cloned for optimal use of this system. Frequently, reactions carried out using this system were characterized by premature termination and the addition of non-template encoded polyuridine tracts to the ends of products. In later work, eukaryotic whole cell or nuclear extracts were used that either contained or were supplemented with RNA polymerases and other accessory factors. The runoff transcripts obtained from these extracts suffered from some combination of poor yields, incorrect initiation, and premature termination.

These technical barriers have been overcome through the use of transcription systems based on the bacteriophages SP6 and T7, which each encode RNA polymerases that are single polypeptide chains. The SP6 system was originally characterized by Butler and Chamberlin, J. Biol. Chem. 257, 5772-5778 (1982), and then used by Melton, et al., Nucleic Acids Res. 12, 7035-7056 (1984), to produce RNA probes of eukaryotic genes. These in vitro synthesized RNAs are superior to nick-translated DNA probes in their ease of synthesis and in their high specific activity. They also are useful for elaborating details about the mechanisms of RNA processing and for providing an efficient means to program in vitro translation.

The T7 RNA polymerase system was first characterized by Studier and co-workers, J. Mol. Biol. 153, 527-544 (1981) and Proc. Natl. Acad. Sci. USA 81, 2035-2039 (1984). This single subunit enzyme has a molecular weight of 92 kilodaltons (kDa), and has been cloned and over-expressed in bacteria to aid in its purification. T7 polymerase is highly specific for a 23 base pair promoter sequence that is repeated seventeen times in the T7 genome, but which has not been found in E. coli or other host DNAs. The viral promoter elements that are required for efficient transcription initiation have been incorporated into high copy vectors with multiple cloning sites for transcription templates, as reported by Rosenberg, et al., Gene 56, 125-135 (1987).

In both the T7 and the SP6 systems, a simple reaction of a few components is sufficient to obtain efficient in vitro synthesis, as described by Milligan, et al., Nucleic Acids Res. 15, 8783-8798 (1987). The T7 system is presently favored because of the greater number of initiations (greater than 100 versus less than 10) obtainable per template molecule as compared to the SP6 polymerase. The T7 RNA polymerase can initiate transcription from a promoter which is as small as eighteen base pairs. The transcribed sequence can be single stranded, so that transcripts up to tRNA length, about 80 nucleotides, can be obtained from a template which has a double stranded promoter and single stranded coding sequence. This system has limitations, because T7 RNA polymerase prefers to initiate transcription at a G and, in addition, the sequence of the transcript from +1 to +6 has a marked effect on the yield of product.

Chemical Synthesis of RNA.

A complementary approach to the in vitro synthesis of RNA is the use of chemical synthesis, as described by Cedergren, et al., Biochem. Cell. Biol. 65, 677-729 (1987). Early workers in this field were stymied by a number of problems, especially the reactivity of the 2' hydroxyl and the relative ease of hydrolysis of RNA under mild alkaline conditions. In order to bring chemical RNA synthesis up to the level of simplicity and repeatability of chemical DNA synthesis, an effective protecting group for the 2' hydroxyl is required, as described by Caruthers, et al., Chem. Scr. 26, 25-30 (1986). Usman, et al., and others, J. Am. Chem. Soc. 109, 7845-7854 (1987) and Biochemistry 28, 2422-2435 (1989), demonstrated the feasibility of the in vitro synthesis of long ribonucleotides by development of 3'-O-phosphoroamidites that were protected at the 2' position with a tert-butyldimethylsilyl (TBDMS) moiety. In conjunction with controlled pore glass supports, the use of these monomers has permitted the complete chemical synthesis of a 77-nucleotide RNA sequence corresponding to tRNA.sup.Met, as reported by Ogilview, et al., Proc. Natl. Acad. Sci. USA 85, 5764-5768 (1988). When tested with a purified preparation of methionine tRNA synthetase, the chemically synthesized tRNA had a methionine acceptance of 11% of that of the native tRNA. The chemical approach provides methods for introducing unusual bases into RNA, mixed intra-chain RNA-DNA hybrid molecules, and other RNAs not available through enzymatic means.

Proteins which Interact with tRNAs and tRNA-Like Structures.

As discussed above, the aminoacyl tRNA synthetases are an ancient class of enzymes that catalyze the two-step aminoacylation reaction. There is one enzyme for each amino acid, and that enzyme charges all isoacceptors of its cognate tRNA species. In the first step of the reaction, the amino acid is activated by condensation with ATP to produce a bound adenylate; subsequently, the activated amino acid is transferred to the 3' end of the cognate tRNA. The esterified tRNA forms a complex with elongation factor Tu, which delivers the charged tRNA to the ribosome. Although all synthetases catalyze the same reaction, they are diverse with respect to sequence, length, and quaternary structure, as reviewed by Schimmel, Ann. Rev. Biochem. 56, 125-158 (1987).

When the first sequences of the tRNA molecules were obtained, the base pairing that gives rise to stems and loops suggested the two-dimensional cloverleaf structure that is now the conventional schematic representation of tRNAs. This was confirmed by extensive physical studies. A complete three-dimensional model of a tRNA, based on x-ray diffraction and crystallographic studies and defining the positions of all nucleotide residues, was first described for yeast phenylalanine tRNA (tRNA.sub.Phe). Kim, et al., Science 185:435 (1974); see also Robertus, Nature 250:546 (1974). At the time of the elucidation of the first sequence, the function of conserved unpaired bases in the cloverleaf was unknown. The x-ray structure revealed the participation of conserved nucleotides, such as U8, A14, G15, G22, G46 and .phi.55, in unusual base-pairing schemes. These interactions feature triple base pairs, reverse Hoogsteen base pairs, and hydrogen bonds between bases and the sugar-phosphate backbone. Collectively, they establish the compact, L-shaped structure of tRNA, whereby the four helical stems are fused into two helices (the acceptor and T.phi.C stems are stacked together, as is the D-stem with the anticodon stem) and the D- and T.phi.C-loops are annealed together. The triple base pair between G22, C13, and G46 in tRNA.sup.Phe strengthens the interaction between the T.phi.C and dihydrouridine loops, and provides greater resistance to thermal, chemical, and enzymatic degradation. Base pairs can also hydrogen bond with the free 2'-hydroxyl of ribose or, as in the ternary interaction between G18, .phi.55, and phosphate 58, with the phosphate oxygen from another portion of the backbone. The specificity of an aminoacyl-tRNA synthetase for its cognate tRNA molecule lies in the three-dimensional structures of the two molecules. The sequence elements that establish the recognition of one tRNA by the aminoacyl-tRNA synthetase has been reported by Schimmel, Biochem. 28:2747 (1989). The G3:U70 base pair in the amino acid acceptor helix is unique to tRNA.sup.Ala and is a major determinant in identifying alanine. Hou & Schimmel, Nature 333:140 (1988); Francklyn and Schimmel, Nature 333:478 (1989); Park et al., Biochem 28:2740 (1989); Hou and Schimmel, Biochem. 28:6800 (1989).

One structural feature demonstrated to be common to several synthetases is that sequences involved in adenylate synthesis are localized to the amino terminal part of the protein, while some of the sequences involved in tRNA binding are located in the carboxyl terminal half. The most conserved structure is the dinucleotide binding fold, an alternating arrangement of beta strands and alpha helices that contains the sequences responsible for adenylate synthesis.

The recognition problem has been investigated for a number of years by many different approaches, as reviewed by Schimmel, Biochemistry 28, 2727-2759 (1989) and Ann. Rev. Biochem. 56, 125-158 (1987), Normanly and Abelson, Ann. Rev. Biochem. 58, 1029-1049 (1989), Schulman and Abelson, Science 240, 1591 (1988), Yarus, Cell 55, 739-741 (1988), Schimmel and Soll, Ann. Rev. Biochem. 48, 601-648 (1979). An important distinction between this interaction and that between regulatory proteins and DNA is that synthetase discrimination between tRNAs can occur at a binding and at a catalytic step. Unlike the interaction of a repressor with a DNA operator, the tRNA-enzyme complex must dissociate quickly to maintain protein synthesis. Consequently, the interaction is not as tight as repressor-operator interactions, and this limits the extent to which recognition can be achieved at the binding step. Dissociation constants at pH 7.5 are in the range of 0.1 to 1.0 .mu.M, which is at least four orders of magnitude weaker than a typical repressor-operator complex. The study of numerous cognate and non-cognate synthetase interactions has shown that, for some complexes, binding discrimination may only contribute a 100-fold preference for the correct tRNAs, while discrimination at the transition state of catalysis may be as high as 104.

In one of the earliest systems for studying tRNA recognition, variants of an E. coli supF amber suppressor (normally inserts tyrosine at UAG codons) were isolated that were aminoacylated with glutamine, Ozeki, et al., Transfer RNA: Biological Aspects (Soll, Abelson, Schimmel, eds) pp. 341-362 (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 1980). Determination of the minimal sequence changes associated with mischarging identified several positions within the acceptor end of the tRNA. These mutations included A73.fwdarw.G, as well as substitutions for the G1:C72 base pair. The molecular basis of glutamine mischarging with these mutant tRNAs was obscure, as some of these changes did not bring the suppressor sequence into closer agreement with tRNA.sup.Gln. In the absence of an understanding of the molecular basis of glutamine mischarging, it was not clear how this type of genetic selection could be applied to other tRNAs. With the tRNA.sup.Gln-GlnRS co-crystal now in hand, the effect of these mutations on mischarging can be rationalized.

Anticodon Substitutions.

Abelson, et al., Proc. Natl. Acad. Sci. USA 83, 6548-6552 (1986), synthesized a set of tRNA genes coding for amber-suppressing tRNAs (CUA anticodon), with the object of defining the minimal set of nucleotide substitutions that are required to convert a tRNA from one amino acid identity to another. So far, introduction of the CUA amber anticodon into 11 of 20 tRNAs does not change the amino acid attached in vivo, as reported by Normanly and Abelson, Ann. Rev. Biochem. 58, 1029-1049 (1989). This set includes: Ala, Gly, Cys, Phe, ProH, HisA, Lys, Ser, Gln, Tyr, and Leu. The remaining tRNAs can be divided into two groups; the first, which includes tRNAs for Ile, Gly, Met, Glu, and Trp, are all mischarged with glutamine. The second group includes those CUA-anticodon tRNAs which are mis-charged with lysine tRNA synthetase (Ile, Arg, Met (m), Asp, Thr, and Val). For those tRNAs that are mischarged when their anticodons are substituted, one or more bases in the anticodon may be a recognition determinant for the cognate enzyme. Additionally or alternatively, it may be a determinant for the glutamine or lysine tRNA synthetases.

"Transplantation Assay".

Those tRNAs unaffected by anticodon changes can be studied through an in vivo "transplantation assay", as devised by Normanly, et al., Nature 321, 213-219 (1986). In this method, base substitutions are introduced into an amber suppressing tRNA gene, which is then transformed into an E. coli strain that also carries a plasmid with a reporter gene that bears an amber (UAG) mutation. If the amber suppressor is functional, the gene product from the reporter gene (typically dihydofolate reductase) is sequenced to determine the identity of the amino acid that has been inserted at the amber codon. By this method, introduction of twelve nucleotides that are common to a set of serine tRNAs into a leucine tRNA isoacceptor was sufficient to confer some serine acceptance in vivo. Since then this approach has been extended to the study of tRNAAla, tRNAPhe, and tRNA.sup.Arg.

Amber suppression is a valuable method for studying how the introduction of nucleotide substitutions into a tRNA sequence affect the amino acid identity of the tRNA. It is restricted, however, to those isoacceptors whose amino acid identities are preserved in the presence of a CUA anticodon. Another problem is that some variants do not accumulate to reasonable intracellular levels, owing to the effect of the nucleotide changes on stability and/or recognition by the processing system. Still another drawback to this approach is that the identity of a tRNA is influenced by competitive reactions between synthetases. Some tRNA variants can act as substrates in vivo for more than one aminoacyl tRNA synthetase. Consequently, altering the levels of synthetases by varying their relative gene dosages will change the amino acid acceptor identity of any "dual identity" tRNA. This phenomenon has been analytically treated by calculations that are based on kinetic parameters for aminoacylation in vitro with alanine and tyrosine of a tRNA.sup.Tyr variant which encodes the major determinant for alanine identity, and is thus charged by tyrosine and alanine. The identity of a tRNA may therefore represent the outcome of many potentially competing interactions between a tRNA and the whole set of cognate and non-cognate synthetases in the cell. For these reasons, examining the interaction of a tRNA with its cognate synthetase in the absence of competing interactions provides information that is obscured by amber suppression.

Since amber suppression can, in some cases, occur with substrate variants which charge poorly or not at all in vitro, suppression can be insensitive to large variations in the efficiency of aminoacylation and cannot be used to make an analytical estimate of the contribution of specific nucleotides to recognition. Studies carried out in vitro circumvent the problems associated with the excess sensitivity of amber suppression, which is influenced by factors in addition to aminoacylation.

The sequence elements that establish the recognition of several tRNAs by the aminoacyl-tRNA synthetase have been reported by Schimmel, Biochem. 28:2747 (1989). These sequence elements were determined by constructing synthetic minihelices corresponding to regions of the targeted RNA and comparing their activity with that of the intact molecule.

As reported by Francklyn and Schimmel, Nature 337:478 (1989), a synthetic minihelix that reproduces the base pairs of the tRNA.sup.Ala amino acid acceptor stem and includes the G3:U70 base pair can be aminoacylated at a rate similar to that of tRNA.sup.Ala, suggesting that this is the primary interaction site between the aminoacyl-tRNA synthetase and the tRNA molecule. Aminoacylation efficiency is markedly improved when the minihelix includes A73. In contrast, minihelices with substitutions at the 3:70 sites are not aminoacylated by alanine tRNA synthetase.

Nucleotide sequence variants of amber-suppressing derivative of E. coli tRNA.sup.Ala have also been screened, focusing on variations on the inside of the L-shaped structure. Two base pairs were found to confer alanine tRNA identity, as reported by Hou & Schimmel, Nature 333:140 (1988).

Evidence indicates that the enzyme finds access to the base pairs in the RNA helix through the minor groove.

Further data supports the utility of this screening method. For example, E. coli supF amber suppressor anticodon was substituted in twenty tRNAs for their anticodons. Recognition of some amino acids by the tRNAs were changed. In another example, E. coli tRNA.sup.Met anticodons were substituted with other anticodons. The results showed that this anticodon is critical to recognition by methionine tRNA synthetase; other regions of sequence have secondary importance for specificity. When the valine anticodon was substituted in tRNA.sup.Met and the methionine anticodon in tRNA.sup.Val, the tRNAs interacted with the opposite synthetase from the original. Similar results were obtained when the arginine anticodon was substituted.

In addition to altering the amino acid recognized by the tRNA, in studies using yeast tRNA.sup.Phe, substituting bases in the anticodon decreased the rate of aminoacylation with the synthetase. Transplantation experiments, discussed below, defined the recognition sequence as consisting of only five bases.

Using an in vivo transplantation assay, twelve tRNA.sup.Ser nucleotides substituted in tRNA.sub.Leu conferred serine binding on the tRNA.sup.Leu. Similar results were obtained with tRNA.sup.Ala, tRNA.sup.Phe, TRN.sup.Arg.

Design of Compounds Targeted to RNA Sequence in Combination with Secondary Structure.

Extrapolated from Comparisons of Protein-Nucleic Acid Interactions.

Once it is understood that RNA has short, specific regions that are critical to its activity, compounds specifically inhibiting the RNA can be designed and synthesized using methodology derived from studies using DNA and DNA-protein interactions, in combination with an understanding of the differences in the chemical and physical composition of RNA as compared to DNA, and knowledge as to the specific region to be inactivated.

The binding of proteins to specific sites in double-stranded DNA is an integral part of gene regulation, DNA synthesis, repair, recombination, and cleavage. X-ray structures have been obtained for several protein-DNA complexes, all of which result from sequence-specific contacts with B-form DNA through major groove interactions. The chemical basis for the discrimination between different base pairs lies in the order of hydrogen bond acceptor and donor groups across the base pair that is accessible to a protein. In principle, this potential array of hydrogen bonds permits all four base pairs to be distinguished from each other on the basis of major groove interactions. In each protein-DNA complex, the conformation of the protein, sometimes in conjunction with bends or kinks in the DNA conformation, acts to position uniquely the specificity-determining polar side chains with respect to the major groove in an orientation that is idiosyncratic to the complex. The nature of the base pair recognized by any particular amino acid side chain will depend on local geometry.

As initially suggested by modeling studies, reported by Lewis, et al., Cold Spring Harbor Symp. Quant. Biol. 47, 435-440 (1983), based on the uncomplexed proteins and helix swapping experiments using repressors such as the lac repressor, reported by Wharton and Ptashne, Cell 38, 361-369 (1984), the repressors use a conserved .alpha.-helix-.beta.-turn-.alpha. helix to contact DNA, with the second of the two helices lying directly in the major groove. Polar side chains in this structural unit make contact with major groove bases in the operator through a series of hydrogen bonds and, occasionally, through hydrophobic interactions. Variations on this basic theme are found. For example, in lambda repressor, the amide NH group of the side chain of Gln44 donates a hydrogen bond to ring N7 in an A:T pair, while the side chain carboxyl oxygen accepts a hydrogen bond from the exocyclic N6 of adenine. This bidentate interaction is further stabilized by a hydrogen bond from the amide group of Gln44 to the amide carboxyl of Gln33, while the amide amino group of Gln33 donates a hydrogen bond to the phosphate oxygen 5' to the A:T pair. Thus, amino acid-base pair contacts can be part of a network of specific hydrogen bonds. In the case of trp repressor, tightly bound water molecules are thought to provide specificity by bridging between groups that are not in direct contact. Hydrogen bonds from peptide amide groups to the phosphate backbone may help to maintain specificity by fixing the orientation of the helix-turn-helix with respect to the major groove. Often, subtle features of the DNA sequence influence the specificity of these protein-DNA interactions by modulating the DNA conformation, so as to create a molecular surface which is complementary to the protein, as discussed by Aggarwal, et al., Science 242, 899 (1988).

Like the repressors, Eco RI endonuclease also uses .alpha.-helices to make hydrogen bonds with the major grooves of its GAATTC recognition sequence, but the recognition helices do not assume a helix-turn-helix structure. The amino-terminal ends of the two recognition helices in each of the two subunits point into the major grooves bases of the inner tetranucleotide AATT. This places specificity-determining amino acid side chains in the proper orientation for base recognition: the carboxyl group of Glu144 receives hydrogen bonds from the successive N6 adenine exocyclic amino groups, while the Arg145 guanidinium donates two hydrogen bonds to the imidazole N7 nitrogens of the adenines located across the axis of symmetry. These "bridging" contacts, in which a single amino acid makes hydrogen bonds to functional groups on two successive base pairs, are unique to the Eco RI complex. The hydrogen bonds donated by each Arg200 guanidinium group to the O6 and N7 of the outer guanines, by contrast, are typical of the contacts made by the repressors.

Given that there are limited restrictions on RNA shape and conformation, there are no simple symmetry considerations that might suggest how proteins recognize RNA sequences. However, the RNA 11 conformation of RNA helices imposes some limits on the potential interactions with protein side chains. In particular, the deep groove of this helical conformation is too narrow for protein structural motifs such as the .alpha. helix to make direct sequence-specific contact. Therefore, the primary basis for sequence discrimination in RNA is believed to be the minor groove. As shown in FIG. 3, there are fewer differences in the pattern of potential hydrogen bond donors and acceptors presented by G:C and A:T (or A:U) base pairs from the face of the minor groove than from the face of the major groove. Because both C and U have the 2-keto group as a potential hydrogen bond acceptor in the minor groove, discrimination between some of the base pairs may be based on the exocyclic 2-amino group of guanine. This expectation is fulfilled in the structure of the Gln-tRNA synthetase-tRNA.sup.Gln complex, reported by Woo, et al., Nature 286, 346-351 (1980).

The three-dimensional structures of transfer RNAs are closely similar. With some specific local features that are idiosyncratic to each tRNA, the molecule features two helical regions, one of which terminates in the amino acid acceptor end, while the other terminates in the anticodon. As a result, the structure of yeast tRNA.sup.Phe can be used as a model for interpreting results on the sequence-specific recognition of most tRNAs. After the yeast tRNA.sup.Phe structure became available, Rich and Schimmel, Nucl. Acids Res. 4, 1649-1665 (1977), considered photochemical cross-linking, tritium labeling, and nuclease digestion data on synthetase-tRNA complexes and proposed that recognition is mediated principally through contacts made along the inside surface of the tRNA "L". On this surface, both helical domains are potential sites for sequence-specific recognition through minor groove discrimination. In addition, at the inside of one end of the "L" the anticodon is a natural site for discrimination because the bases are unpaired, and because this sequence codes for the attached amino acid. On the outside of the L, an alternative region is the "variable pocket", which is formed by the interaction of the T.phi.C and D loops, described by Ladner, et al., Proc. Natl. Acad. Sci. USA 72, 4414-4418 (1975) and McClain, et al., Science 241, 1804-1807 (1988). The nucleotides which comprise this patch, 16, 17, 59 and 60, are not conserved among tRNAs, and are not engaged in Watson-Crick base pairs. Accordingly, several different regions potentially can contribute recognition determinants, and possible interactions include hydrogen bonds to either the minor groove exocyclic amino or keto groups, or to the unpaired bases themselves.

In the co-crystal between E. coli tRNA.sup.Gln and the glutamine tRNA synthetase, the protein binds along the inside of the L-shaped structure. The anticodon and specific acceptor stem nucleotides are in contact with the synthetase. In the acceptor stem, the exocyclic 2-amino group of G2 forms hydrogen bonds to the backbone carboxyloxygen of Pro181 and to the backbone amide of Ile183. The latter interaction is bridged through a bound water molecule, in a fashion reminiscent of "indirect readout" first suggested in the trp repressor complex. A hydrogen bond to the exocyclic 2-amino group of G3 is made by the carboxyl of Asp235, which also hydrogen bonds to the previously mentioned water molecule.

A more complex feature is the interaction of the protein with the 3' end of the acceptor stem, and the conformational change by the nucleotides that are located in the 3' acceptor end. The U1:A72 base pair at the end of the acceptor stem is wedged open by the side chain of Leu136, which protrudes from a .beta.-turn in the acceptor binding domain of the protein. The rate of charging of tRNA.sup.Gln variants is influenced by the propensity of this base pair to be melted out, as described by Seong, et al., J. Biol. Chem. 246, 6504-6508 (1989). The unpaired nucleotides (GCCA76) at the 3' acceptor end are folded back at a angle with respect to the acceptor stem helix, such that the 3' end is buried deep within the dinucleotide binding fold, in close proximity to bound ATP and, presumably, the bound amino acid. The 2-amino group of G73 hydrogen bonds to the phosphate oxygen of the previous nucleotide. This interaction stabilizes the unusual conformation of the acceptor arm, and specifically depends on having a G at position 73. At present, it is clear that the recognition of tRNA.sup.Gln involves, at a minimum, contacts to the exocyclic amino groups in the minor groove and sequence-dependent conformation changes in the tRNA itself.

Comparison of Aminoacylation by Viral RNAs and tRNAs.

Some aminoacyl tRNA synthetases aminoacylate the 3' ends of certain genomic and subgenomic plant viral RNAs. This suggests a structural relationship between tRNAs and the 3' ends of these viral RNAs. Computer predictions of structure were tested experimentally with chemical probes by Dumas, et al., J. Biomolec. Struc. & Cyn. 4, 707-728 (1987), and led to the proposal of an RNA pseudoknot that enables a tRNA-like structure to form at the 3' end, by van Belkim, et al., Nucl. Acids Res. 16, 1931-1950 (1988).

In the RNA pseudoknot, bases in a hairpin loop form Watson-Crick pairs with bases that are located outside of the hairpin structure. Because less than 11 base pairs form with the loop, there is only partial revolution of one strand about the other, so that a true knot is avoided. In the pseudoknot described for turnip yellow mosaic virus, there is co-axial stacking of the two different helical stems of the pseudoknot. The stems are joined by two different connecting loops which cross the major and minor grooves, respectively. The pseudoknot structure is supported by the 2-D NMR studies of Puglisi, et al., Nature 331, 283-286 (1988), on short synthetic RNA fragments, where the stability of the pseudoknot has been shown to be sensitive to temperature and Mg2+ concentration.

Brome mosaic virus (BMV) RNA (aminoacylated by tyrosine) and turnip yellow mosaic virus (TYMV) RNA (aminoacylated with valine) are the most extensively studied plant viral RNAs. For BMV, a synthetic 135-nt fragment retains aminoacylation function. Dreher, et al., Nature 311, 171-175 (1984) explored the sequence requirements for aminoacylation with tyrosine and for viral replication by introducing mutations into the viral 3' end and at a putative AUA "anticodon" sequence. Those substitutions in which the CCA end was changed had abolished aminoacylation function, but retained at least partial replication function. The sequences at the AUA anticodon, by contrast, were not required for aminoacylation, but severely attenuated replication. The genetic separation of aminoacylation and replication functions in BMV RNA suggests that aminoacylation is not required for virus viability.

Dreher, et al., Biochimie 70, 1719-1727 (1988), synthesized a series of length variants of TYMV RNA in vitro and determined kinetic parameters for their aminocylation by wheat germ valine tRNA synthetase. Although 83 3'-terminal nucleotides can be folded into a tRNA-like structure that can be aminocylated in vitro, sequences which lie upstream of this structure (between 82 and 159 from the 3' end) are required for a maximal rate and extent of aminoacylation. The decreased rate of aminoacylation of fragments shorter than 159 nucleotides is reflected predominantly in a decreased Vmax rather than Km, suggesting that the sequences 82-159 affect catalytic rather than binding steps. This is another demonstration of the significance of the transition state for catalysis for recognition by synthetases. Footprinting studies carried out on the fragments with purified synthetase suggest that the enzyme either contacts this region directly, or that this region is required for the correct conformation of the tRNA-like domain. Unlike the BMV RNA, the TYMV "anticodon" is an important determinant for aminocylation, as it is for E. coli tRNA.sup.Val.

Studies Using RNase P, an Ribonucleoprotein Containing an RNA Molecule with Enzymatic Activity Cleaving Pre-tRNAs.

RNase P is required for maturation of the 5' ends of tRNA precursors. The enzyme has two different subunits in all organisms investigated so far. In E. coli, these consist of a 13.7 kDa protein component (C5), and a 377-nucleotide RNA component known as the M1 subunit, as reported by Altman, et al., Trends Biol. Sci. 11, 515-518 (1986). This nuclease distinguishes tRNA precursors from all other RNAs. Mutational analyses of precursor molecules showed that substitutions that disrupt the secondary or tertiary structure of the precursor inhibit the cleavage reaction, by Smith Brookhaven Symp. Biol. 20, 1902-1906 (1981) and McClain and Seidman, Nature (London) 257, 106-(1975). Thus, the enzyme is sensitive to the structure of the precursor. RNA synthesis of enzyme and substrate component has proved to be an effective way to approach recognition of tRNA precursors.

The essential role of RNA in the catalytic event was first demonstrated when cleavage of the precursor tRNA was shown to be dependent on both M1 RNA and C5 protein by Kole and Altman, Biochemistry 20, 1902-1906 (1981) and Reed, et al., Cell 30, 627-630 (1982). Subsequently, Guerrier-Takada, et al., Cell 35, 849-857 (1983) showed that the requirement for C5 could be overcome by raising the Mg2+ concentration from 10 to 60 mM. Kinetic parameters at 60 mM Mg2+ were determined for the holoenzyme reaction and for the reaction with M1 RNA alone. Under these conditions, C5 increased the velocity of the reaction by two-fold, but had no effect on the Km. The rnp A gene that codes for the C5 protein subunit is essential for viability in E. coli, so the operational Mg2+ concentration in vivo maybe closer to that (10 mM) used in the original assays. In vitro, it is possible to carry out complementation experiments utilizing the E. coli RNA and B. subtilis C5 protein. Thus, the protein may recognize features of the RNA structure which have been conserved during evolution.

The C5 protein and M1 RNA components of RNaseP have been cloned and over-expressed. Utilizing these reagents, Vioque, et al., J. Mol. Biol. 202, 835-848 (1988), measured a dissociation constant of 4.times.10.sup.-10 M for the binding of M1 to C5. Footprint analysis showed protection of nucleotides between 82-86 and 170-270. A competition assay was used to examine the binding of synthetic truncated derivatives of M1 RNA to C5. A fragment formed of sequences from 94 to 272 effectively competed away binding to non-specific RNAs, while a fragment spanning either 1-168 or 164-272 did not.

A phylogenetic comparison of nine different sequences from two different eubacterial phyla established a "consensus" RNase P (Min 1) that contained only 263 nucleotides versus the 354 to 417 nucleotides of the parental structures, and incorporated stems, loops, and pseudoknot features that were conserved between all members of the collection, as reported by Waugh, et al., Science 244, 1569-1571 (1989). The Min 1 consensus also contained one of the regions implicated by footprinting, i.e., the sequences between 82 to 96 in E. coli M1 RNA. In vitro transcripts of the Min 1 structure processed a pre-tRN.sup.Asp substrate at a rate that was only five-fold slower than that of full length E. coli M1 RNA. The success of this design strategy is consistent with the belief that particular structural determinants of RNase P have been conserved through evolution.

The region from 86 to 92 in M1 has been further implicated by enzyme-substrate cross-linking studies by Guerrier-Takada, et al., Science 38, 219-224 (1984). Mixtures of M1 RNA and a pre-tRNA.sup.Tyr were irradiated with UV light at 300 or 254 nm, and then resolved on polyacrylamide gels to isolate the specific complexes. Reverse transcriptase was used to establish the points of crosslinking in both the enzyme and substrate. Reverse transcription terminated consistently at C93 in M1 RNA, indicating that C92 is cross-linked to the substrate. The cross-linking experiments also defined points of contact to the substrate. Efficient termination of reverse transcription (using a primer complementary to the 3' end of the tRNA precursor substrate) occurred at G-2, two nucleotides before the start of the mature tRNA. This indicated that C92 in M1 is cross-linked to "C-3" in the pre-tRNA. This is within three bases of the cleavage site in the pre-tRNA.

Deletion of C92 in M1 RNA raised Km by 100-fold and lowered kcat by 6-fold relative to wild type M1, in the absence of C5. However, the specific nucleotide at position 92 is not critical, because a U92 substitution mutant had nearly the same kinetic parameters for processing as wild type M1. Also, deletion of C92 can be partially overcome by the presence of the C5 subunit. Thus, N92 may influence the local conformation of the RNase P active site, but may be secondary to the influence of the C5 protein subunit.

In parallel with the work on the M1 RNA, in vitro RNA synthesis has also been used to investigate the substrate requirements for the reaction. Truncated version of E. coli tRNA.sup.Phe that retain the acceptor-T.phi.C stem and loop are substrates for the enzyme, but the introduction of base substitutions at C74 (the 3' terminus is A76) eliminated cleavage. As described for alanine tRNA synthetase, RNase P recognizes a limited part of the overall tRNA structure. There is also evidence to suggest that RNase P recognizes the 3' CCA sequence of the precursor tRNA molecule, as reported by Guerrier-Takada, et al., Cell (1984). The precursor to E. coli tRNA.sup.Tyr is three nucleotides longer at the 3' end than the mature species, such that the sequence is CCAUCAOH. Cleavage of this substrate in vitro with M1 RNA or the RNase P holoenzyme reveals that the turnover number for the reaction with M1 RNA alone is greatly reduced in the absence of the CCA sequence. The wild type M1 RNA will correctly cleave a pre-tRNA.sup.Tyr which lacks the 3' terminal CCAUCA, although at a slower rate than for the wild type precursor. A mutant RNase P with a deletion of C92 also cleaves the mutant precursor, but does so at a site that is 4 to 6 bases upstream of the wild type cleavage site. Reverse transcription of a photo-crosslinked complex between the mutant M1 RNA and the mutant pre-tRNA.sup.Tyr gave strong termination at G1 in pre-tRNA. Since high concentrations of exogenous CCA trinucleotide inhibit the reaction of a substrate that contains the CCA group, but stimulates the processing of a substrate that lacks the trinucleotide, RNase P may have two separate binding sites for the pre-tRNA, one associated with the eventual site of cleavage, and one for the CCA end.

In the case of synthetic tRNAs, one drawback is that the transcripts are unmodified. The lysidine in the E. coli tRNA.sup.Ile2 isoacceptor is one example of a modified base shown to be essential for aminoacylation with isoleucine. However, unmodified transcripts have been used to purify and characterize several nucleotide modification enzymes, including pseudouridine synthase from S. cerivisiae, and guanine methyltransferase from Xenopus oocytes. Microinjection of in vitro transcripts into Xenopus oocytes can be used to produce modified tRNAs in vivo. As more of the genes coding for the tRNA modification enzymes are cloned and their gene products characterized, tRNA transcripts produced in vitro can be treated with these enzymes to study the effects of modifications on molecular recognition.

Claim 1 of 13 Claims

1. A method for designing a compound specifically inhibiting targeted ribonucleic acid function comprising the steps of: (a) determining the nucleotide sequence in the targeted ribonucleic acid that is critical to function; (b) determining the secondary structure of the region of the targeted ribonucleic acid in which the critical site is located; (c) determining the three-dimensional stature of the targeted RNA, including the position of the critical site relative to the major and minor grooves; (d) determining the sequence of nucleotides and structure flanking the critical site in the targeted ribonucleic acid that is specific to the critical region of the ribonucleic acid to be inhibited and within the minor groove; and (e) synthesizing a compound that will bind specifically to the critical site within the minor groove of the targeted ribonucleic acid thereby inhibiting targeted ribonucleic acid function.


If you want to learn more about this patent, please go directly to the U.S. Patent and Trademark Office Web site to access the full patent.



[ Outsourcing Guide ] [ Cont. Education ] [ Software/Reports ] [ Training Courses ]
[ Web Seminars ] [ Jobs ] [ Consultants ] [ Buyer's Guide ] [ Advertiser Info ]

[ Home ] [ Pharm Patents / Licensing ] [ Pharm News ] [ Federal Register ]
[ Pharm Stocks ] [ FDA Links ] [ FDA Warning Letters ] [ FDA Doc/cGMP ]
[ Pharm/Biotech Events ] [ Newsletter Subscription ] [ Web Links ] [ Suggestions ]
[ Site Map ]